Page 104 - Vitamin D and Cancer

4 The Epidemiology of Vitamin D and Cancer Risk 91

groups. A limitation of the study was the relatively small number of total cancers
of 50 in total and 37 after the first year. Nonetheless, total cancer incidence was
lower in the Ca + D women than in the placebo-control subjects (P < 0.03), and the
RR of incident cancer was 0.40 (P = 0.01) in the Ca + D group and 0.53 (P = 0.06)
in the Ca-only group. In a sub-analysis confined to cancers diagnosed after the first
year, the RR for the Ca + D group was 0.23 (95% CI = 0.09–0.60; P < 0.005); no
significant risk reduction was observed for the Ca-only group. In multivariate models,
both vitamin D treatment and higher 25(OH)D levels were significant, independent
predictors of reduced cancer incidence.

4.11 Summary

Ecologic studies that compared cancer mortality rates in different regions within
the USA initiated the hypothesis that high vitamin D levels may lower risk of vari-
ous cancers, a hypothesis that was subsequently supported by biologic evidence.
Colorectal cancer was the first cancer type hypothesized to be related to vitamin D
status [3]. Subsequently, although regional UV-B was shown to be associated with
a number of cancers, the magnitude of the association appeared to be strongest for
colorectal cancer [8]. This finding for colorectal cancer was confirmed in epidemio-
logic studies of circulating 25(OH)D and colorectal cancer risk, in which individu-
als in the high quartile or quintile of 25(OH)D had a 40–50% risk reduction relative
to those in the lowest group. Inverse associations have also been observed for pre-
dicted vitamin D, sun light exposure and dietary intake, and for the colorectal
cancer precursor, the adenoma. The consistency of this association in diverse cir-
cumstances indicates that an uncontrolled or unaccounted confounding factor is
unlikely to account for these associations.
For breast cancer, for which an inverse correlation has been observed between
regionally estimated UV-B in ecologic studies, the epidemiologic data are sparser
and less consistent. The evidence from analytic epidemiologic studies of vitamin D
and breast cancer are somewhat conflicting. There have been only two adequately
powered prospective studies of circulating 25(OH)D levels, and these have yielded
inconsistent results. The studies of vitamin D intake are modestly supportive but
limited by the generally low intakes of vitamin D. One case–control study provided
intriguing findings: more sun exposure primarily during ages 10–19, but not at
other ages, than controls, was inversely associated with risk of breast cancer.
Because recall bias is a possible explanation, replicating these results in prospective
settings is important. Interestingly, adolescent exposures have often been found to
be critical in determining subsequent breast cancer risk, probably because the
breast tissue are rapidly developing over this time period.
Ecologic studies of regional UV-B and cancer mortality find an inverse associa-
tion with prostate cancer mortality. However, this association appears not as strong
as that for colorectal or breast cancer [8], and in one study, was limited to counties
north of 40°N latitude, in the USA [88]. The studies of circulating 25(OH)D have